$Roundhill Memory ETF(DRAM)$ Saw a slight pullback earlier. I'm planning to reload and add some more heading into the weekend. For now I'm just taking the opportunity and accepting the volatility. I think higher highs are ahead in due time.
$Roundhill Memory ETF(DRAM)$ It's become pretty clear to me that South Korea as a whole hasn't really learned anything from the market crash. I'm still adding more DRAM but definitely selling calls and buying puts this time around. DRAM hitting 100+ one day feels inevitable. It is going to be a bumpy ride though.
Hopefully the US tech recovery we're seeing provides some relief to KOSPI, which is down around 20% after five straight red sessions. DRAM is 47% Samsung and SK Hynix, and those two make up more than half of KOSPI, so the chain is pretty direct. I think we need Korean investors to stabilize and start buying again before $Roundhill Memory ETF(DRAM)$ can realistically get back to the $60, $70, and $80 range. $Micron Technology(MU)$ and $SanDisk Corp.(SNDK)$ aren't going to carry this thing on their own.
$NEBIUS(NBIS)$ $Roundhill Memory ETF(DRAM)$ China isn't slowing down on AI expansion — they're accelerating. It's hard to imagine the US wouldn't respond in kind and keep pushing its own buildout. At this point, it doesn't really feel like there's much of a choice.
$Roundhill Memory ETF(DRAM)$ I know this price level is painful for a lot of people. Honestly, stepping away from checking DRAM and memory holdings for a while might save you some unnecessary stress. This probably gets worse before it gets better, which is just part of investing.
$Roundhill Memory ETF(DRAM)$ First, people said AI demand is peaking. Then, the narrative shifted to China flooding the market with cheap memory. Meanwhile, global AI and data center infrastructure spending is projected to approach $7–7.6 trillion by 2030. That scale of buildout doesn't happen without massive demand for GPUs, HBM, DRAM, networking, storage, and power. The long-term AI thesis hasn't changed.
$Roundhill Memory ETF(DRAM)$ The recent drop came in faster and more brutal than expected, but do we really believe the AI story is over? If not, memory is going to be a big part of that growth. And if that's the case, then this isn't some pump and dump, even though the price action makes it feel like one right now. Just have to let the SK Hynix earnings, the Fed, and capex news play out this week.
$Roundhill Memory ETF(DRAM)$ The future is going nowhere without memory chips, it's that simple. DRAM is being heavily manipulated today along with the rest of the chip space, but it's just manipulation.
$Roundhill Memory ETF(DRAM)$ It looks like big money really wrecked this one. I suppose the bottleneck was somehow magically resolved in just a quarter. Amazing.
$Roundhill Memory ETF(DRAM)$ The underlying companies are generating strong cash flow and could realistically deliver 2-5x returns from these levels by the time the capex cycle wraps up. I can understand the skepticism around thematic ETFs packed with unprofitable names, but memory and storage companies have genuinely solid fundamentals backing them.
$iShares Semiconductor ETF(SOXX)$ $Roundhill Memory ETF(DRAM)$ $Defiance AI and Power Infrastructure ETF(AIPO)$ I've been using Claude Code pretty much all day for professional work, and my subscription jumped from $20/month to $200/month in just two months. Honestly, it's easily worth it — saves a ton of time and makes me sharper than I'd be on my own. I don't even feel like a programmer anymore. More like I'm managing a genius employee who works almost for free. So yeah, infrastructure spending is going to keep going and keep growing. No doubt about it.
One thing that stands out is this has been outperforming $Micron Technology(MU)$ for the first time in weeks. The agreements with Nvidia seem to have had an effect. If $Invesco QQQ(QQQ)$ flips green, $Roundhill Memory ETF(DRAM)$ could really move. There is also earnings this week. Feels like this might be one of the last chances to buy the dip.
$Roundhill Memory ETF(DRAM)$ Current DDR5 DRAM on 10nm is fully booked for the next three years. DDR6 on 5-8nm is expected to hit the market around 2027, and it will drive an industry-wide refresh since it doubles speed while cutting power consumption. DDR6 and DDR7 will also be used in space systems for the obvious power, weight, and thermal advantages. Memory has a 5-7 year runway from where things stand now. Only a few players in the game.
$Roundhill Memory ETF(DRAM)$ If hyperscalers keep pouring money into AI, memory stocks look like the main beneficiaries and will probably do well. GOOGL was down while memory names were up, so there's that divergence worth noting. Curious to see if that strength holds over time.
$Roundhill Memory ETF(DRAM)$ It's almost making up for the rough stretch last month. My calls are up 86%. I've set a trailing stop. I'm just going to relax, have a drink, and watch how this plays out.